{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/iterative-amortized-inference","title":"Iterative Amortized Inference","arxiv_id":"1807.09356","date":"2018-07-24","proceeding":"ICML 2018 7","authors":["Joseph Marino","Yisong Yue","Stephan Mandt"],"abstract":"Inference models are a key component in scaling variational inference to deep\nlatent variable models, most notably as encoder networks in variational\nauto-encoders (VAEs). By replacing conventional optimization-based inference\nwith a learned model, inference is amortized over data examples and therefore\nmore computationally efficient. However, standard inference models are\nrestricted to direct mappings from data to approximate posterior estimates. The\nfailure of these models to reach fully optimized approximate posterior\nestimates results in an amortization gap. We aim toward closing this gap by\nproposing iterative inference models, which learn to perform inference\noptimization through repeatedly encoding gradients. Our approach generalizes\nstandard inference models in VAEs and provides insight into several empirical\nfindings, including top-down inference techniques. We demonstrate the inference\noptimization capabilities of iterative inference models and show that they\noutperform standard inference models on several benchmark data sets of images\nand text.","url_abs":"http://arxiv.org/abs/1807.09356v1","url_pdf":"http://arxiv.org/pdf/1807.09356v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"iterative-amortized-inference","repo_url":"https://github.com/joelouismarino/iterative_inference","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"inference-optimization","task_name":"Inference Optimization"},{"task_slug":"variational-inference","task_name":"Variational Inference"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1807.09356","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}